Reasons to Participate: The Role of Political Attitudes in Digital Activism

In this paper, we study the effects of political attitudes on digital activism. We operationalise political attitudes through three constructs used in classical political science studies: citizenship norms, institutional trust and psychological political engagement. In addition, following the social-psychological models of explanation of political conduct, we suggest the existence of a qualitative difference between the affective component of attitudes, which would coincide with psychological political engagement, and the cognitive component of attitudes, which are made up of political beliefs and values. The latter influences the former, which, in turn, has an influence on political conduct. To achieve our goals, we have analysed digital political participation in Spain. Our results support the qualitative differentiation between cognitive and affective components of attitudes; as a result, there is a clear need to study those political beliefs that specifically characterise digital activism.


Introduction
This paper focuses on one of the emerging research areas in regard to the relationship between politics and the Internet: digital activism (Bennett & Segerberg, 2013;Castells, 2015). 1 We define digital activism, or digital political participation, as the repertoire of participatory forms carried out by citizens through the Internet (Robles, Antino, De Marco, & Torres, 2015). According to the literature, this form of political participation is, first, related to intrinsic issues of the digital field, such as Internet users' electronic resources (Gibson, Lusoli, & Ward, 2005;Ausserhofer & Maireder, 2014) and digital skills (van Deursen & van Dijk, 2009;van Dijk & van Deursen, 2014). Second, the influence of political factors, such as the ideology and offline political participation of Internet users, has been studied (van der Meer, van Deth, & Scheepers, 2009;Calenda & Mosca, 2007;Harlow & Harp, 2012;Robles, De Marco, & Antino, 2013). This paper aims to gain a better understanding of how the political attitudes of Internet users influence digital political participation practices.
For decades, general literature on offline political participation has looked at the role of political attitudes (Barrett & Brunton-Smit, 2014;Labigne, 2012). There are two great traditions: the classical approach, based on political, scientific and sociological perspective, and the socialpsychological traditions. Unlike a social-psychological approach, classical models do not emphasise distinctions between different facets of the political attitude construct. According to classical models, there are three different kinds of political attitudes: citizenship norms, institutional trust and psychological political engagement (Abramson & Aldrich, 1982;Niemi, Craig, & Mattei, 1991;Sniderman & Grob, 1996;Verba, Schlozman, & Brady, 1995). Each one of these would have the same kind of impact on political participation.
On the other hand, social-psychological models make a qualitative distinction between cognitive and affective components of attitudes (Ajzen, 1991;Caprara & Zimbardo, 2004): political beliefs and values on the one hand and engagement in politics on the other. This distinction implies a qualitative differentiation between citizenship norms and institutional trust, considered as the cognitive component of attitudes, and psychological political engagement, which represents the affective component. Furthermore, authors conceive the influence of cognitive components of political attitudes over political action, that is, citizenship norms and institutional trust, also taking into account the potential indirect effect through psychological political engagement, considering the affective component of political attitudes as the very precursor of political action (Cohen, Vigoda, & Samorly, 2001;Caprara & Zimbardo, 2004).
Traditional research has mainly focused on studying the influence that political attitudes play on both institutional and non-institutional political participation. These studies have contributed much evidence of this type of relationship (Abramson & Aldrich, 1982;Verba, Schlozman, & Brady, 1995). However, literature on digital political participation has studied, in a less systematic way, the influence of political attitudes on online forms of political behaviour (Best & Kruger, 2005). This is due to the rise in academic interest in studying determinant factors of digital political participation, compared with the older tradition of empirical research on the effects of digital political participation on the political system (Delli & Carpini, 2000;Margolis & Resnick, 2000;Bimber, 2001;Krueger, 2002;Ward et al., 2003;Tolbert & McNeal, 2003;Quintelier & Vissers, 2008;Schlozman et al., 2010;Bonilla & Rosa, 2015).
The first goal of this paper is to analyse, in a more systematic way, the relationship between political attitudes and digital political participation. To do this, a representative set of political attitudinal indicators, especially of citizenship norms, institutional trust and psychological political engagement, will be taken into consideration. The paper also attempts to apply the knowledge generated by classical and social-psychological models regarding the relationship between political attitudes and political participation within a context that addresses the political use of the Internet. We put both theoretical references to the test to understand our subject of research. This is to analyse which model is better at predicting the behaviour defined as digital political participation. Our aim therefore is twofold; we shall try to understand whether the classical attitudinal variables studied in political science and sociology allow us to predict digital political participation and whether these relationships reveal the hierarchies suggested by a social-psychological perspective.
If the results support the social-psychological hypothesis, then this work would enable further reflection on those beliefs and values that specifically influence digital political participation. This is an interesting area of research, and one that is gaining growing relevance within the field of political sociology (Norris, 2001;Best & Kruger, 2005;Borge & Cardenal, 2011;Robles, Antino, De Marco, & Torres, 2015).
To meet our research goals, we analyse digital political participation in Spain, more specifically, data from the "Internet and political participation" study (ref. 2736), undertaken by the Centre for Sociological Research (Centro de Investigaciones Sociológicas, CIS). Using path analysis, we shall test whether political attitudes have an effect on digital political participation. At the same time, we shall try to understand whether psychological political engagement plays a role in the relationship between citizenship norms, institutional trust and digital political participation, which would confirm the suggestions of social-psychological models.

Political Attitudes and Digital Political Participation: Classical Models and Psychosocial Models
As previously mentioned, political attitudes have been studied, from the point of view of the classical approach, through three attitudinal dimensions: citizenship norms, institutional trust and psychological political engagement. Dalton (2008) defines citizenship norms as "the whole number of shared expectations for the role of citizens in politics" (p. 78). The literature points to three types of citizenship norms. The first, based on the elitist model, is defined on the basis of respecting the norms (Almond & Verba, 1963). The second is based on deliberation and criticism (Kymlicka & Norman, 2000), and the third on solidarity (Conover, Searing, & Crewe, 2002). However, according to Denters, Gabriel and Torcal (2007), on an individual level, these dimensions are not appreciable, and the construct of citizenship norms could be considered as a compact block.
Institutional trust has been defined as a citizen's belief as to whether political institutions fulfil their role, even when citizens are not permanently supervising them (Citrin & Muste, 1999). This theoretical construct would therefore be a reflection of the general feelings of citizens regarding their country's institutions (Newton & Norris, 2000) and, more specifically, their trust in them. In turn, trust is the result of the idea that each political institution is honourable, credible, competent, transparent and efficient (Levi & Stoker, 2000).
Lastly, psychological political engagement mainly refers to citizens' interest in politics and public affairs (Milbrath, 1977). It is also related to the citizen participation in conversations about politics, as well as with efforts to persuade others of one's own political viewpoint (Verba, Schlozman, & Brady, 1995). The empirical literature highlights that this attitudinal construct has the greatest influence on both institutional and non-institutional political participation (Orum, 1989;Verba, Nie, & Kim, 1987). For this reason, this construct plays a particularly relevant role when it comes to analysing the political participation of citizens.
Studies that are undertaken from the classical approach show that all these variables, when considered individually, have a direct influence on political participation (Bolzendahl & Coffe, 2013;Hooghe & Marien, 2013;Dalton, 2013). However, the majority of these studies have only focused on one attitudinal dimension at a time, avoiding a joint assessment of the three political attitudes. This would limit the understanding of whether these attitudes are interrelated and in which way they impact political participation. That is, this approach does not allow an understanding of whether they are part of the same construct or whether each of the attitudinal dimensions represents a construct in themselves.
Then there are the social-psychological models (Cohen, et al., 2001;Valencia, 1990;Bekkers, 2005). These models are concerned with integrating individual constructs characteristic of psychosocial research with elements of research topics from the political sociology field. More specifically, the individual indicators would coincide with political attitudes. An attitude can be defined as the evaluation of an object that motivates the subject's action and is influenced by the individual's beliefs and values (Ajzen, 1991;Eagly & Chaiken, 1993;Fishbein & Ajzen, 1975;Petty, Wegener, & Fabrigar, 1997). According to Ajzen (1991), attitudes are a multidimensional construct composed by cognitive, affective and behavioural components. On one hand, beliefs and values are the foundations of the cognitive component of political attitudes (Ajzen, 1991).
Beliefs, in general, reflect the different visions of political realities that surround subjects (Conover & Fieldman, 1984). Specifically, beliefs about political institutions, as we have just mentioned, would result from an evaluation of the extent to which those public institutions that represent citizens are respectable and honourable (Citrin & Muste, 1999;Bonet, Martín, & Montero, 2006;Newton & Norris, 2000). Values, for their part, are the citizens' views regarding just what is politically desirable. These values serve as guides in the political life of an individual (Rohan, 2000). Studies concerning political values have mostly focused on the idea of a "good citizen" (Almond & Verba, 1963;Kymlicka & Norman, 2000;Conover, Searing, & Crewe, 2002). Thus, these values guide the subjects' political actions, encouraging them to act in line with their citizenship ideals.
According to Caprara and Zimbardo (2004), this distinction would also apply to political attitudes, as beliefs and values would form part of the cognitive component of this attitude. On the other hand, the affective component of political attitudes corresponds with psychological political engagement, that is, with the emotional and prebehavioural part of attitude. This psychological engagement is considered to be a favourable disposition towards politics and can be seen as the stronger predictor of the behavioural component of attitudes (Ajzen, 1991). This dimension would be moulded by the subject's interest in politics and by their intention to take part in the political process through specific actions.
Cognitive components of attitudes would change depending on the typology of political participation considered (Zukin et al., 2006). In fact, it is a reasonable assumption that people who are more inclined to noninstitutional participation, such as demonstrations and boycotting, would have different political values and beliefs from people who prefer institutional forms of participation, such as voting or participating in the activities of political parties. Meanwhile, psychological engagement would represent the invariable core of political attitudes (Ajzen & Fishbein, 2000;Caprara & Zimbardo, 2004). Both institutional and non-institutional activists would have the same level of psychological engagement.
There are studies to support this thesis; according to Theocharis (2011), post-materialist values, hegemonic among Greek youths, are closely linked to new forms of political participation (Robles, De Marco, Antino, & Lobera, 2016). Similarly, these values help us predict digital activism. In fact, this argument is not new. Norris (2001) points out that this is a process of change in political practices as a result of longer-term social change that is closely related to a change in the type of prevailing social values.
Literature that study the relationship between political attitudes and digital activism need more empirical studies and more specific research (van der Meer, van Deth, & Scheepers, 2009;Robles, Antino, De Marco, & Torres, 2015). This lack is mainly due to the fact that research on digital political participation has been organized according to the debate between those who hold that digital political participation may foster offline political participation by the mobilisation of a section of the inactive population (Delli Carpini, 2000;Ward et al., 2003;Krueger, 2002;Tolbert & McNeal, 2003;Quintelier & Vissers, 2008;Loader, B. & Mercea, 2011;Ausserhofer & Maireder, 2014) and those who hold that the Internet has no effect on offline political participation (Margolis & Resnick, 2000;Bimber, 2001;Schlozman, Verba, & Brady, 2010). As a matter of fact, academics that have studied the effects of digital political participation on the political system have generally seen digital political participation as an independent variable. However, a small number of studies have taken into account those factors that exclusively influence digital activism (Livingstone et al., 2005;De Marco, Robles, & Antino, 2014). From this perspective, digital political participation is considered to be a dependent variable, and specialists have been focused on studying its explanatory factors.
As with all other studies on political behaviour, studies on digital political participation often use a set of control variables that measure the resources of the respondents, thus allowing an analysis of the effect of aspects such as age, educational level and ideology. Specific studies have been undertaken to examine the commonality of using the Internet for political purposes among young people and among those with higher educational levels (Calenda & Meijer, 2009). In most cases, the use of the Internet for political purposes is more frequently distributed among people with left-wing ideology (van der Meer et al., 2009). However, the relationship between attitudes and digital political participation was really not taken into account by social-psychological nor classical traditions. Those that did, only took a small number of political attitudes indicators into consideration, those basically related to an interest in politics (Best & Kruger, 2005).
Our aim in this paper is, first, to test the influence of all the classical attitudinal constructs, that is, citizenship norms, institutional trust and psychological political engagement, on digital political participation. This would allow us to undertake an in-depth study of the relationship between political attitudes and digital political participation. Second, we aim to uncover whether these constructs can be interpreted from a social-psychological perspective.
As such, we shall first test the effect of the three classical political attitude variables on digital political participation. Second, we shall verify whether institutional trust and citizenship norms have an indirect effect on digital political participation through psychological political engagement, such as the existence of a difference between cognitive and affective components of attitudes seems to suggest. Finally, we will discuss the implications of our analysis for future empirical studies on the relationship between political attitudes and digital political participation. Our idea is to emphasise the need to focus our studies on the kinds of values and beliefs that represent the cognitive component of political attitudes, which will better help us to predict digital political participation modes.
If our analysis provides empirical support for a socialpsychological perspective, we should then look at the kinds of beliefs and values that are most useful to predicting an individual's political use of the Internet. The social-psychological thesis considers the central role of classical political beliefs and values in explaining political participation. This consideration is being extended to digital political participation (Borge & Cardenal, 2011). However, beliefs and values are not static; for example, some authors have explained how new social movements are active agents in creating social and political realities by generating new identities, cognitions, beliefs and values (Laraña et al., 1994;Melucci, 2001;Micheletti, 2003). Norris (2001) began to discuss the relationship between social change and the development of such new forms of participation as digital political participation. In the last section of this paper, we return to this debate because our empirical research points to an interest in studying new values and beliefs as a variable for explaining digital political participation.

Data and Sample
To meet our goals, we have used the CIS study "Internet y participación política" (Internet and Political Participation) (ref. 2736). This survey is the first in Spain to compile information about the penetration of the Internet in Spain and its various uses (including political uses) by the Spanish population as well as indicators on political attitudes. The survey has a sample of 3,716 subjects referring to a universe composed of the Spanish population of both sexes. 2 More specifically, 1,829 men and 1,886 women (50.8% of the sample) have been included in the study. Additionally, the sample includes the variable "age" (33.39 of the sample). Finally, for the purposes of this research, we have decided to remove those subjects from the study that did not know what a personal computer was (Q. 18) and those that had not connected to the Internet in the three months prior to the survey (Q. 19); thus, a sample of 2,169 subjects was obtained that is exclusively composed of Internet users.

Independent Variables
With regard to citizenship norms, institutional trust and psychological political engagement, three indices have been used comprising the 18 indicators that were used in Study 2736, which was introduced to measure these classic attitudinal constructs. These indices are the result of a prior factorial analysis (De Marco, Antino, & Robles, 2012) 3 which has allowed us to extract three dimensions, perfectly coinciding with the three variables subject to study, 4 and which have been called citizenship norms, institutional trust and psychological political engagement, respectively. (The three dimensions were computed through the factorial scores three, obtaining three variables with standardized scores.)

Variables Dimension
Level of trust in the central government.

Institutional Trust
Level of trust in political parties. Level of trust in trade unions. Level of trust in councils. Level of trust in the media. Level of trust in NGOs.
To be a good citizen, how important is it to comply with the law?

Citizenship Norms
To be a good citizen, how important is it to not avoid paying taxes? To be a good citizen, how important is it to vote? To be a good citizen, how important is it to think of others? To be a good citizen, how important is it to form your own personal opinion?
Interest in politics.

Psychological Political Engagement
How often do you speak about politics? How often do you try to persuade other people of your own point of view? I am better informed about politics than others. Politics is difficult to understand. Using the Internet to contact a politician or political party.
Using the Internet to contact an administration to complain or protest.
Using the Internet to communicate with an association or organization.
Using the Internet to donate or raise funds for a campaign or for an organization/association. Writing comments on a blog or forum with social or political content.
Signing an online petition. Tables 1a and 1b shows the variables included in each of the dimensions extracted by the factorial analysis. The items used to measure institutional trust all used scales with intervals ranging from 0 to 10, with 0 representing the minimum confidence value and 10 the maximum value. The items that make up the "citizenship norms" dimension were assessed using Likert scales, measuring the level of agreement with different statements, where 1 represented "strongly disagree" and 5 represented "strongly agree". The other indicators used Likert scales from 1 to 4, except for an indicator that measures the frequency with which subjects' try to convince those close to them of a political idea, whose scale ranges from 0 to 5, and an indicator measuring the frequency with which they talk about politics, whose scale ranges from 0 to 6.
We have also included control variables such as ideology (numeric variable, with intervals from 0 to 10, where 0 represents "extreme left" and 10 represents "extreme right") and level of education. The last one is an ordinal variable, with 7 response categories that represent different years of education. All of these variables, according to different empirical studies, are considered to be possible sources of influence on digital political participation (Norris, 2001;Best & Kruger, 2005; van der Meer, van Deth, & Scheepers, 2009).

Dependent variable
Study 2736 includes six dichotomous items to measure digital political participation (Anduiza et al., 2010). We have included them in Table 2. Our dependent variable has been constructed by adding up all the affirmative replies to each item. In this way, a numeric variable has been obtained (with values from 0 to 6 included) which quantifies the number of different digital political activities carried out by the subjects surveyed.

Data Analysis
A path analysis model has been implemented to understand whether cognitive components of attitudes have an indirect effect on digital political participation through the emotive component of attitude, as suggested by several psycho-sociological models for offline participation. We then attempted to understand whether the effect of the variables assessing institutional trust and citizenship norms on digital political participation presented an indirect effect through psychological political engagement. To verify these sets of direct and indirect effects, we used a path analysis approach.
The reason for choosing this technique is that it offers the possibility to gauge the fit of our theoretical   models to our empirical data. We run a first model with institutional trust, citizenship norms and psychological political engagement (correlated) on digital political participation (Model 1 in Figure 1), controlling for levels of education and political ideology. We then run a second model (Model 2 in Figure 1) where effects of the variables assessing institutional trust and citizenship norms (correlated) on digital political participation were presented an indirect effect through psychological political engagement; specifically, we compared the direct effects of institutional trust and citizenship norms on digital political participation (Model 1) with the indirect effects of institutional trust and citizenship norms on digital political participation through psychological political engagement. Table 2 presents the descriptive analysis (mean and standard deviation) of our variables as well as the bivariate correlations amongst them. In the same table, we included the reliability estimated through Cronbach's Alpha calculated on the items included in the factorial analysis.

Path Analysis
As shown in Table 3, an individual's ideology maintains significant influence across all dimensions that make up the construct of political attitudes. However, an individual's level of education only has an effect on political psychological engagement.
As to the study of the influence of institutional trust and citizenship norms on digital political participation, we found that the direct effects of institutional trust and citizenship norms on digital political participation were not significant in the first model. As shown in Table 3, in the second model, we found a significant effect of institutional trust and citizenship norms on political psychological engagement, as well as a significant relationship between political psychological engagement and digital political participation.
Moreover, the results shown in Tables 3 and 4 indicate that Model 2 presents a good fit with the theoretical model (χ 2 /df = 5.129, TLI = 0.93, CFI = 0.98, RMSEA = 0.044, SRMR = 0.021), while Model 1 had an unacceptable fit. We additionally ran a chi square comparison between the two models (∆χ 2 = 95.09, ∆df = 1, p < .01), and we found that Model 1 presented a significantly better fit to the data. These results provide empirical support for the existence of an indirect effect, specifically on the effects that the institutional trust and citizenship norms variables correlated have on digital political participation through political psychological engagement.

Discussion
Spain is a country where the use of the Internet has undergone a significant development over the past decade. Currently, the Internet penetration rate in Spain is 78.7%. In political terms, thanks to a survey undertaken by the Autonomous University of Madrid (UAM-2014) on attitudes and the digital divide, devised in 2012 and consisting of 1,500 interviewees, representative of the Spanish population, we know that a quarter of the population of Internet users has undertaken, at the very least, one political activity over the Internet, such as contacting a politician online, participating in social protest events actions or sharing political content on a social network. Thanks to the same survey, we know that more than half of those surveyed believe that the Internet improves their ability to influence power (55.5%). This result is very important because it shows that most Spaniards believe that the Internet can empower them politically.
In this context, our work informs us about political attitudes of digital activists in Spain. This type of political behaviour, which is becoming increasingly widespread nationally, requires a set of psychological dispositions that  are associated with the affective and cognitive components of attitudes toward politics. This finding, although expected, is new to the literature on digital political participation; although, of course, it is not as new within the scope of political participation offline. However, our study allows us to go a step further in understanding how attitudinal variables and digital political participation are related. We now know that not all attitudinal components have the same kind of effect on this type of digital behaviour, as previous literature about the relationship between political attitudes and offline participation suggested (Abramson & Aldrich, 1982;Niemi, Craig, & Mattei, 1991;Sniderman & Grob, 1996;Verba, et al., 1995). Thus, psychological political engagement functions as the chain of transmission triggered between cognitive components and digital political participation. All this is thanks to our ability to combine two traditions in the study of political attitudes: the classical and the social-psychological traditions.
Taking a more in-depth look at this, our research allows us to point to, first, the importance of already tested control variables, such as educational attainment, when it comes to predicting digital political participation (Best & Krueger, 2005;Norris, 2001). We know that the more educated citizens are, the more likely they are to engage in digital political practices. There is speculation around the idea that educated citizens have more digital resources at their disposal to carry out practices such as making political use of the Internet, which require medium-high capabilities of handling the tool. Also, data showed significative relationships between control variables and attitudinal variables. More specifically, the analysis showed that right-wing ideology is positively related with citizenship norms, and negatively related with institutional trust and psychological political engagement. At the same time, education has a positive relationship with all the attitudinal dimensions: increasing the number of years in the educational system has a positive impact on institutional trust, citizenship norms and psychological political engagement.
Furthermore, the path analysis model shows the positive influence that attitudinal dimensions have on digital political participation, in mirroring what occurs with offline forms of participation, being psychological political engagement a relevant variable in predicting digital political participation. Finally, as the socialpsychological model holds, the results of our path analysis show that there is a qualitative difference between values and political beliefs on the one hand, that is, citizenship norms and institutional trust, and the affective component of the attitudes, that is, psychological political engagement, on the other. This is a first step towards the exploration of values and beliefs that also influence psychological political engagement but which characterize digital political participation more exclusively.
The results shown have a double implication on the study of digital political participation. First of all, once the relationship between political attitudes and online participation has been established, it opens up the necessity to study the way in which attitudinal constructs may foster digital political participation. The second important implication of our results, the concatenation of influences between cognitive and affective components of attitudes and digital political participation, create some relevant questions about the possibility that digital political participation is related to specific values and beliefs, that is, to specific cognitive components of political attitudes. To date, only the political values and beliefs linked to offline political participation, such as citizenship norms and institutional trust, have been explored. However, we believe that an important line of research should focus on the beliefs associated with new information and communication technologies and their political potential. Likewise, some authors suggest a change in the forms of political participation in Western countries, characterised by a decrease in the traditional forms of participation, and, at the same time, by a diversification of the forms of citizen engagement in political affairs (Hague & Loader, 1999).
Recently, a great variety of creative forms of participation have emerged, thus doing away with the traditional distinction between private and public life (Micheletti & McFarland, 2011). Digital political participation is included among these. Thus, new types of values and beliefs remain to be explored which, although not related to the traditional forms of participation, will be related to the new forms of participation like digital political participation (De Marco, Antino, & Robles, 2012;Robles, De Marco, & Antino, 2013).
Our body of research points to the need for future work to include research and surveys regarding the factors of influence of digital political participation, indicators of post-materialist values and beliefs, such as the freedom of expression, tolerance of diversity, social egalitarianism, secularism, internationalism and democratic participation (Norris, 2001). The inclusion of these indicators would allow an in-depth study of the characteristics of digital activists.
This work becomes a perfect complement to other recently published works by the authors that focus on the factors that explain the effect of beliefs about political possibilities of technologies on digital political participation (Robles, Antino, De Marco, & Torres, 2015). This paper highlights how such beliefs, greatly determined by the individual's perception on the ease of use of technologies, are a key factor in understanding digital political participation. In this way, we believe we are making progress in understanding how different types of beliefs, considered as cognitive components of attitudes, and the political use of the Internet are interrelated.

Limitations
One of the most important limitations of our study is that the data used for this research were collected eight years ago. It is reasonable to assume that digital political participation is a behaviour characterized by rapid evolution, since technology is developing at a fast rate, offering Internet users new spaces and instruments to communicate, interact and participate. This means that our research has only looked at a snapshot of digital political participation in Spain at a specific time. At present, this is the only available data that studies this phenomenon with a representative sample (in both Spain and the other southern European countries), as confirmed by the absence of empirical evidence in this regard.
A second limitation concerns the measurement of the construct of digital political participation, which fails to include more innovative forms of digital political participation, such as those related to social networks. We are conscious that this issue affects the measurement of the phenomenon (since those new practices are not included), but we have also determined that it does not invalidate the whole measurement. There are several reasons to consider that these new behaviours are positively related with other digital political participation behaviours, such as those included in this study. If all digital political participation behaviours are positively related, the measurement of a limited set of digital political participation behaviours will still provide a valid estimation of digital political participation, although the inclusion of these new practices would reduce any measurement errors.

Notes
1 This paper is part of research project CSO2009-13771, funded by the Ministry of Science and Innovation of the Spanish Government. 2 Multi-stage sampling, stratified by conglomerates.
Non-proportional allocation. The use of the sample in general has involved the use of weighting. There are three populations of reference: National, Andalusian and young people between the ages of 18 and 44. For a confidence level of 95.5% (two sigmas), and P = Q, actual error rate is ±1.64% for the entire sample and in the case of simple random sampling (Source: Technical data sheet from Study 2736 by CIS). 3 We based our choice on a previous analysis that offers empirical evidence for construct validity. (De Marco, Antino, & Robles, 2012). Specifically, we first conducted an exploratory factorial analysis (using the WLS extraction method and Promin rotation). The obtained structure was interpreted according to the theoretical background (also proposed in this paper), offering evidence of construct validity. Although our factors do not cover the whole spectrum of psychological concepts of political beliefs and values, we consider that the included indicators offer a valid measurement of those constructs (see Torcal & Montero, 2006). Finally, we calculated individual's scores of the three extracted factors with the regression method. 4 We excluded the construct political self-efficacy based on a factorial analysis that excluded a large part of the items in CIS's Study 2736 that measure this variable.